||Our objective is to estimate geothermal reservoir indicators, such as: P and S seismic velocity models to a depth of ~300 m, ambient noise spectral energy and media stochastic properties. An important advantage of our method is estimating the shear velocity model, which, unlike the P-velocity model, is not yet accomplished by conventional reflection surveys. We analyze ambient seismic noise recorded by a 3 day, 1.3 km2, 100 m spaced vertical geophone survey as well as four 12 m-separation seismic lines. The survey was conducted by UNR and Imageair Inc. in March 2010 at Soda Lake, Nevada, geothermal field operated by Magma Energy Corporation. We use seismic interferometry, a new imaging method, to generate subsurface images without any larger seismic sources, such as explosions or earthquakes. One application of seismic interferometry is to retrieve the impulse response or Green’s Function (GF) from crosscorrelation of ambient seismic noise. The ambient-noise autocorrelation at each station is interpreted as the collocated source–receiver elastic wave Green’s Function (i.e. the Earth’s reflection response). Stacks of ambient noise crosscorrelations at pairs of sensors over three days result in inter-station GF’s, with Rayleigh waves as dominant arrivals. A preliminary estimation of the velocity of phases which we interpret as fundamental Rayleigh waves shows lower surface wave velocity and higher scattering within the geothermal production field, at frequencies of 1-5 Hz. Using array processing techniques, such as frequency-wavelength (fk) analysis, we will estimate Rayleighwave phase velocity dispersion curves. The dispersion estimates will be inverted for surface wave velocity models using the Computer Programs in Seismology (CPS3.0) surf96 algorithm. Stacks of autocorrelations of ambient noise data recorded at individual sensors result in retrieval of the Earth’s reflection response at the location of each sensor. The autocorrelation traces are interpreted in terms of reflection GF phase composition and crustal reflector properties. By applying crosscorrelation to ambient noise data recorded at pairs of sensors located 12 m apart we generate virtual shot gathers as if one of the sensors is generating seismic waves, i.e. we retrieve the Earth’s reflection response. We will also investigate whether differences between production and non-production geothermal reservoir areas could be assessed by measuring seismic scattering. We will compare the stochastic parameters (Hurst number, characteristic length) from the ambient noise autocorrelations and crosscorrelations and the ambient noise spectral energy differences above the geothermal reservoir to similar parameters outside the geothermal reservoir area.